Abstract
Vitamin A (retinol) is an important nutrient for embryonic development and adult health. Early studies identified retinoic acid (RA) as a metabolite of retinol, however, its importance was not apparent. Later, it was observed that RA treatment of vertebrate embryos had teratogenic effects on limb development. Subsequently, the discovery of nuclear RA receptors (RARs) revealed that RA controls gene expression directly at the transcriptional level through a process referred to as RA signaling. This important discovery led to further studies demonstrating that RA and RARs are required for normal embryonic development. The determination of RA function during normal development has been challenging as RA gain-of-function studies often lead to conclusions about normal development that conflict with RAR or RA loss-of-function studies. However, genetic loss-of-function studies have identified direct target genes of endogenous RA/RAR that are required for normal development of specific tissues. Thus, genetic loss-of-function studies that eliminate RARs or RA-generating enzymes have been instrumental in revealing that RA signaling is required for normal early development of many organs and tissues, including the hindbrain, posterior body axis, somites, spinal cord, forelimbs, heart, and eye.
Introduction
The history of retinoic acid (RA) signaling began with the discovery that its precursor vitamin A (also called retinol) is essential for embryonic growth and development as shown by vitamin A deficiency studies (Wilson et al. 1953). Other early studies demonstrated that RA is a naturally occurring metabolite of retinol that can be generated by alcohol dehydrogenase that metabolizes retinol to retinaldehyde (Bliss 1951), followed by aldehyde dehydrogenase metabolizing retinaldehyde to RA (Elder & Topper 1962). Although a normal function for RA remained elusive for many years, it was observed in the 1970s and early 1980s that vertebrate embryos treated with RA suffered limb teratogenesis (Kochhar 1973, Maden 1982, Tickle et al. 1982), suggesting that RA may be important for limb development.
A major breakthrough in how RA controls development occurred upon discovery of nuclear RA receptors (RARs), which use RA as a ligand to control transcription of key genes (Giguère et al. 1987, Petkovich et al. 1987). Humans and mice possess three RARs (RARA, RARB, and RARG) that are required for many developmental processes as shown by RAR double knockout studies which were needed to overcome redundant functions among the three RARs (Lohnes et al. 1994, Mendelsohn et al. 1994). RARs bind to RA target genes as a heterodimer complex with one of the three retinoid X receptors (RXRA, RXRB, or RXRG) (Kastner et al. 1995). Binding of RAR/RXR occurs at what is known as an RA response element (RARE) that consists of two 6 bp repeated DNA sequences (Hoffmann et al. 1990). RAR/RXR bound to a RARE can recruit nuclear receptor coactivators (NCOA) or nuclear receptor corepressors (NCOR) depending upon binding of RA to the RAR portion of the heterodimer, thus directly activating or repressing transcription of nearby genes (Kumar et al. 2016). A ligand for RXR is not required for RA signaling during the early stages of embryogenesis (Mic et al. 2003).
Discovery of the requirement for RARs in normal development stimulated studies designed to identify enzymes needed to generate RA from its natural source, vitamin A (retinol) (Fig. 1). RA is produced from retinol in a two-step process as alluded to above in early metabolic studies. In embryos, RA synthesis was found to be initiated by retinol dehydrogenase-10 (RDH10) that converts retinol to retinaldehyde in specific tissues (Sandell et al. 2007). To prevent excessive production of retinaldehyde, reverse metabolism of retinaldehyde back to retinol is performed by dehydrogenase/reductase-SDR family member 3 (DHRS3) (Feng et al. 2010, Billings et al. 2013); interestingly, DHRS3 physically interacts with RDH10 to form a complex that regulates the rate of RA synthesis by modifying the retinol/retinaldehyde ratio (Belyaeva et al. 2017). In the second step of RA synthesis, retinaldehyde is metabolized to RA by three retinaldehyde dehydrogenases expressed in specific tissues, i.e., ALDH1A1, ALDH1A2, and ALDH1A3 (also known as RALDH1, RALDH2, and RALDH3) (Niederreither et al. 1999, Mic et al. 2002, Dupé et al. 2003, Fan et al. 2003). Conversion of retinaldehyde to RA is irreversible. Thus, in order to eliminate an excessive amount of RA, three P450 family enzymes (CYP26A1, CYP26B1, and CYP26C1) degrade RA in specific tissues; this results in a short (~1 h) half-life for RA (Hernandez et al. 2007, Pennimpede et al. 2010). It is important to note that although retinol is available in all cells of the embryo, RA synthesis and RA degradation occur only in specific tissues to enable tissue-specific RA signaling. For instance, ALDH1A2 first appears in development at E7.5 expressed only in presomitic mesoderm, but at E8.5 ALDH1A2 is also expressed in the optic vesicles and ALDH1A3 begins expression in the optic vesicles at E8.5, with ALDH1A1 beginning expression in the optic vesicles at E9.5 (Sirbu et al. 2005, Molotkov et al. 2006). Also, CYP26 enzymes are expressed in specific locations, such as certain hindbrain rhombomeres, the caudal progenitor region, and the distal limb, resulting in the degradation of RA in specific locations (Abu-Abed et al. 2002). Thus, during early development, RA is generated in specific tissues, diffuses to nearby tissues, and is degraded in specific tissues, resulting in the presence of RA only in certain tissues and hence the ability of RA signaling to control the development of specific tissues.
Mechanism of retinoic acid (RA) synthesis, RA degradation, and RA signaling. Retinol is transported in the blood via retinol-binding protein-4 (RBP4) which makes it available to all cells of the body. In embryos, retinol is metabolized to the intermediate metabolite retinaldehyde by retinol dehydrogenase-10 (RDH10) and then retinaldehyde is metabolized to RA by one of three aldehyde dehydrogenase-1A (ALDH1A) enzymes. Retinaldehyde can also be converted back to retinol by dehydrogenase/reductase-SDR family member 3 (DHRS3). RA can be degraded by one of three CYP26 enzymes to limit RA activity. RA can act locally or be released from cells in which it is generated and travel as a signal to other cells. RA enters the nucleus and binds to one of three RA receptors (RARs) that form a heterodimer complex with one of three retinoid X receptors (RXRs). RAR/RXR heterodimers bind a repeated DNA sequence known as an RA response element (RARE) whose consensus sequence is AGGTCA (N1,2,5) AGGTCA where N refers to 1, 2, or 5 random nucleotides. RA directly controls transcription by serving as a ligand for RARs. Binding of RA to RARs results in a conformational shift in RAR that alters its ability to bind nuclear receptor coactivators (NCOA) or nuclear receptor corepressors (NCOR) that control further downstream events in transcriptional activation or repression.
Citation: Journal of Molecular Endocrinology 69, 4; 10.1530/JME-22-0041
Genetic loss-of-function studies have been instrumental in determining the normal functions of RA signaling during development (Rhinn & Dolle 2012, Cunningham & Duester 2015, Ghyselinck & Duester 2019). RAR double knockout studies revealed many functions of RA signaling during the late stages of embryogenesis, but these studies did not reveal the early embryonic functions of RA signaling due to the presence of one remaining RAR (Lohnes et al. 1994, Mendelsohn et al. 1994). Fortunately, Aldh1a2 single knockouts and Rdh10 single knockouts have provided a wealth of information on the earliest functions of RA signaling during embryogenesis. Also, Cyp26 knockouts revealed abnormal development due to the presence of RA signaling in locations that do not normally have such signaling (Pennimpede et al. 2010). Studies on these knockouts revealed that embryonic RA signaling begins just a few hours before the formation of the first somite (vertebral progenitor) and is limited to the developing eye (optic vesicle) and posterior body axis (hindbrain, spinal cord, and trunk) (Sirbu et al. 2005, Cunningham et al. 2015) (Fig. 2). This early somite stage is just a snapshot of what is happening around E8.5. As development proceeds to later stages, the locations of RA signaling change as the locations of Aldh1a and Cyp26 expression change. Here, we will focus on the early functions of RA signaling during development.
Retinoic acid (RA) signaling target tissues during early embryogenesis. Early mouse embryos that have 1–15 somites exhibit RA signaling activity (shown in green) in the developing eye (optic vesicle) and throughout the posterior body axis (developing trunk); the anterior border of RA signaling activity in the trunk occurs in the middle of the hindbrain, and the posterior border occurs at the junction with the caudal epiblast that contains neuromesodermal progenitors (NMPs) that differentiate to form either spinal cord or somites; the diagram shows a dorsal view of a 7-somite mouse embryo. Tissues that require RA signaling for normal development include the optic vesicle, posterior hindbrain, somites, spinal cord, forelimbs, NMPs, and heart that is located underneath on the ventral side of the embryo.
Citation: Journal of Molecular Endocrinology 69, 4; 10.1530/JME-22-0041
Role of RA during limb development
As previously mentioned, one of the earliest reported actions of RA is its teratogenic ability to cause defects in limb anteroposterior patterning (thumb to little finger) and proximodistal limb patterning (upper arm to hand) when vertebrate embryos are treated with pharmacological levels of RA (Maden 1982, Tickle et al. 1982). Subsequent studies showed that RA treatment of distal limb tissue alters limb proximodistal patterning by causing distal expression of normally proximal-specific homeobox genes Meis1 and Meis2 which inhibits distal limb development normally controlled by distal Fgf8 expression (Mercader et al. 2000); these studies also showed that FGF8 treatment of proximal limb tissue downregulates Meis1/Meis2. Later studies demonstrated that RA normally generated in the trunk by Aldh1a2 diffuses into proximal limb tissue (Niederreither et al. 1999, Mic et al. 2002). Additionally, Cyp26b1 is expressed in distal limb tissue and degrades distal RA for normal limb proximodistal patterning (Yashiro et al. 2004), plus Fgf8 was shown to be required for Cyp26b1 expression in distal limb (Probst et al. 2011). Together, these findings suggested that a proximal-high gradient of RA signaling is needed to activate proximal Meis1/Meis2 expression, along with a distal-high gradient of FGF signaling that is needed to repress Meis1/Meis2 distally to enable normal limb proximodistal patterning, thus a two-signal model. Fgf8 knockout studies supported the hypothesis that FGF8 signaling is required for limb proximodistal patterning (Mariani et al. 2008). However, the hypothesis that RA signaling is required for limb proximodistal patterning (and Meis1/Meis2 expression) was not supported by loss of RA signaling in proximal limb tissue performed either with the Aldh1a2 knockout eliminating the second step of RA synthesis (Zhao et al. 2009) or the Rdh10 knockout eliminating the first step of RA synthesis (Cunningham et al. 2013); also, these knockouts demonstrated that RA signaling is not required for limb anteroposterior patterning. Instead, the Aldh1a2 and Rdh10 knockouts demonstrated that RA signaling is required for initiation of forelimb (but not hindlimb) budding; see below for description of this function. Together, all of these studies are the best demonstration that one cannot determine the normal functions of endogenous RA by using RA treatment (gain-of-function) studies, thus pointing out the necessity of performing RA loss-of-function studies (preferably using genetic knockouts) to determine the normal functions of endogenous RA signaling.
The disparate conclusions obtained with RA treatment vs RA synthesis knockout can be explained by side effects due to the high amount of administered RA that ectopically activates Meis1/Meis2 in distal limb (Cunningham & Duester 2015). Based upon the knockout studies one can visualize a one-signal model for limb proximodistal patterning that is driven by distal FGF8 that activates genes needed for limb outgrowth in a proximal to distal direction and that activates Cyp26b1 distally to eliminate RA in distal limb which might activate unwanted distal Meis1/Meis2 expression. Interestingly, although Meis1/Meis2 expression in proximal limb does not require RA, it has been observed that Meis1/Meis2 do have nearby RAREs and their expression in the trunk does require RA for a normal level of expression (Berenguer et al. 2020). These observations suggest that RA treatment can force a gene to be expressed in an ectopic location by swamping its associated RAREs with RAR/RXR bound to RA, thus activating the gene in a tissue that is not normally exposed to endogenous RA.
Role of RA signaling during forelimb initiation
As mentioned above, genetic loss of Aldh1a2 or Rdh10 in mouse does not alter limb proximodistal or anteroposterior patterning, but loss of RA signaling with these mutants prevents initiation of forelimb development, although hindlimb development is not affected. Aldh1a2 mutants lack expression of Tbx5 in trunk lateral plate mesoderm that generates the forelimb field (Zhao et al. 2009), and Rdh10 mutants exhibit greatly reduced Tbx5 expression in the forelimb field (Cunningham et al. 2013). As Tbx5 is essential for forelimb initiation (Agarwal et al. 2003, Rallis et al. 2003), RA functions upstream of this important regulator of forelimb bud initiation. Other studies support a requirement for RA in forelimb bud initiation, i.e., treatment of chick embryos with an RA synthesis inhibitor (Stratford et al. 1996) and vitamin A-deficient rat embryos (White et al. 1998).
Comparison of the forelimb defects in mouse Aldh1a2 and Rdh10 knockouts revealed that RA signaling throughout the trunk anteroposterior axis is required to restrict Fgf8 expression to two domains on either side of the forelimb field (Cunningham & Duester 2015). One domain is located in the anterior trunk where the heart is forming, and another domain is located in the posterior-most region of the body axis (caudal epiblast or tailbud). Loss of RA synthesis results in ectopic Fgf8 expression that stretches posteriorly from the heart into the forelimb field and anteriorly from the caudal epiblast into the forelimb field, associated with loss of forelimb field Tbx5 expression. These findings, plus the observation that WT mouse embryos treated with FGF8 fail to activate forelimb Tbx5, demonstrate that the forelimb bud initiation defects in RA-deficient mutants are due to excessive FGF8 activity in the trunk leading to a disruption of forelimb Tbx5 activation (Cunningham et al. 2013). This permissive model of RA action is supported by studies in zebrafish showing that forelimb bud initiation can be rescued in aldh1a2 mutants by introducing a heat-shock inducible dominant-negative FGF receptor (Cunningham et al. 2013). It is possible that RA may also function instructively to activate Tbx5 or some other gene in the forelimb field, but there are currently no genetic loss-of-function studies that support this model.
RA controls posterior body axis extension and somitogenesis
Bipotential neuromesodermal progenitors (NMPs) are cells in the caudal epiblast of vertebrate embryos that express both T (Brachyury) and Sox2; these progenitor cells undergo balanced differentiation to either presomitic mesoderm (expressing T and Tbx6) or spinal cord neuroectoderm (expressing Sox2 and Sox1) (Henrique et al. 2015). Although RA is not required for NMP establishment (Cunningham et al. 2016), loss of RA results in altered NMP differentiation with decreased Sox1 and Sox2 expression in cells that remain in the epithelial layer (prospective neural progenitors); loss of RA signaling also leads to increased Tbx6 expression in cells that migrate through the primitive streak via the gastrulation process to form mesodermal progenitors resulting in a small-somite defect (Cunningham et al. 2015).
In Aldh1a2 knockout embryos that completely lack RA activity, Tbx6 expression in presomitic mesoderm is increased; however, trunk somites are approximately half of their normal size, suggesting RA is required for proper condensation of migratory fibroblast-like presomitic mesodermal cells to form epithelial somites (Cunningham et al. 2015). Treatment of mouse Aldh1a2 knockouts with the FGF inhibitor SU5402 rescues trunk somite size, suggesting that the mechanism of RA action during somitogenesis is similar to that during forelimb initiation in which RA signaling represses caudal Fgf8; thus, excessive caudal FGF8 signaling evidently interferes with somitogenesis when Fgf8 is expressed too far anteriorly (Cunningham et al. 2015).
The mechanism of caudal Fgf8 repression requires a RARE located 4.1 kb upstream of Fgf8 that is required for caudal Fgf8 repression; this RARE is conserved in amniote embryos such as human, mouse, and chick but is not present near zebrafish fgf8 (Kumar & Duester 2014, Kumar et al. 2016). Interestingly, zebrafish embryos do not require RA for NMP differentiation or caudal fgf8 repression as loss of RA signaling does not affect body axis extension or somitogenesis (Berenguer et al. 2018).
The Fgf8 RARE functions in an RAR/RXR- and RA-dependent manner to recruit NCORs (NCOR1 and NCOR2) and polycomb repressive complex 2 (PRC2) that stimulate deposition of the repressive H3K27me3 chromatin mark near Fgf8 (Kumar & Duester 2014, Kumar et al. 2016). These observations provided the first in vivo evidence that RA signaling can directly repress a gene through a RARE silencer, plus these studies also demonstrated that NCOR can function ligand-dependently for gene repression.
RA regulates early heart patterning
In vertebrates, loss of RA results in a failure of heart tube looping resulting in defective formation of ventricular (anterior) and atrial (posterior) chambers (Niederreither et al. 2001, Keegan et al. 2005). The mechanism of RA signaling in the early heart involves RA-FGF8 antagonism as described above for somitogenesis and forelimb initiation (Ryckebusch et al. 2008, Sirbu et al. 2008, Sorrell & Waxman 2011). FGF8 signaling located in the anterior heart field is needed to establish ventricular identity (Pradhan et al. 2017). As loss of RA signaling results in posterior expansion of heat Fgf8 expression, RA represses Fgf8 to prevent expansion of anterior ventricular progenitors into the posterior region of the heart where atria develop. Thus, proper formation of forelimbs, somites, and heart all require RA repression of Fgf8, pointing out a major function of RA signaling during development (Cunningham & Duester 2015).
RA is required for hindbrain anteroposterior patterning
In mouse embryos undergoing the late stage of gastrulation, RA synthesis begins upon initial expression of Rdh10 and Aldh1a2 in the presomitic mesoderm; RA generated in the presomitic mesoderm is secreted and diffuses into the developing hindbrain as far anteriorly as rhombomere 3 (r3); the hindbrain initially forms eight rhombomeres (r1–r8) located along the anteroposterior axis (Sirbu et al. 2005). In the hindbrain, RA activates expression of the 3’-Hox genes (Hox1-4 groups) that are known to be essential for rhombomere formation and identity (Krumlauf 1993, Maden et al. 1996, Niederreither et al. 2000, Begemann et al. 2001). In the case of Hoxb1, RA signaling directly regulates expression through two RAREs, one that functions as an enhancer to activate expression in r4 and another RARE that functions as a silencer to repress expression in r3 and r5 (Marshall et al. 1994, Studer et al. 1994). Hnf1b (vhnf1), a repressor of Hoxb1, is activated by RA signaling in posterior hindbrain (as far anterior as r5) and spinal cord which also functions to repress Hoxb1 posterior to r4 (Hernandez et al. 2004, Sirbu et al. 2005).
RA control of spinal cord development
As mentioned above, bipotential NMPs undergo balanced differentiation to either spinal cord neuroectoderm or presomitic mesoderm in vertebrate embryos. Loss of RA in mouse or chick embryos results in decreased expression of Sox1/Sox2-expressing neural progenitors which delays differentiation of NMPs to spinal cord neuroectoderm (neural plate) (Cunningham et al. 2015, Henrique et al. 2015).
As the spinal cord further develops, RA also plays a role in differentiation of specific neural progenitors to motor neurons. After neural progenitors emerge from the caudal epiblast to form the spinal cord, they are exposed to RA derived from adjacent somites as well as Sonic hedgehog (SHH) generated in the floor plate of the ventral spinal cord. Together, RA and SHH activate Pax6 and Olig2 expression in ventral spinal cord progenitors to stimulate a motor neuron fate (Diez del Corral et al. 2003, Novitch et al. 2003, Molotkova et al. 2005, England et al. 2011). ChIP-seq studies comparing WT and Aldh1a2 knockouts have shown that no RARE is associated with activation of Pax6; instead, the requirement of RA signaling for Pax6 activation may proceed through several indirect mechanisms due to the ability of RA to directly activate Sox2 and Cdx1, both required for Pax6 expression, along with the ability of RA to directly repress caudal Fgf8 that prevents Pax6 expression in the caudal epiblast (Berenguer et al. 2020).
Function of RA during eye development
Eye development begins when the optic vesicles develop as out-pocketings of the forebrain, and soon afterward, the optic vesicles invaginate to form optic cups with a connection to the forebrain known as the optic stalk that will later form the optic nerve. Early RAR double knockout studies showed that although the optic vesicle forms and undergoes optic cup formation, RA is required afterward for further optic cup morphogenesis and proper formation of anterior eye structures adjacent to the optic cup such as cornea and eyelids (Lohnes et al. 1994); this topic is discussed in more detail below. In order to determine whether RA plays an earlier role in optic cup formation, RAR triple knockout studies could be performed (not yet reported), but for now one can rely on targeting RA-generating enzymes. During the optic vesicle stage, all three ALDH1A RA-generating enzymes are expressed (Molotkov et al. 2006). This triple redundancy initially hampered genetic loss-of-function studies to interrogate eye RA function; however, the discovery that Aldh1a1−/− mice survive as adults and are fertile (Fan et al. 2003) enabled the generation of Aldh1a1/Aldh1a2/Aldh1a3 triple knockout embryos that lack all RA signaling activity in the early optic field. Analysis of such triple knockouts revealed that although the optic vesicle forms, it does not properly invaginate to form the optic cup resulting in loss of the ventral portion of the optic cup (Molotkov et al. 2006).
Subsequent to optic cup formation, RA signaling is also required for eye morphogenetic processes that result in proper formation of anterior eye structures (cornea and eyelids) adjacent to the optic cup (retina). This later function of eye RA signaling was initially discovered by examination of RAR double knockouts (Lohnes et al. 1994) and also by analysis of the Rara/Rxra double knockout, thus revealing a requirement for RXR in eye morphogenesis (Kastner et al. 1994). During the late stage of eye RA signaling, RA is generated only by Aldh1a1 (optic cup dorsal retina) and Aldh1a3 (optic cup ventral retina) since expression of Aldh1a2 has ceased in the optic field (Molotkov et al. 2006). Although Aldh1a1 and Aldh1a3 single knockouts revealed no obvious eye defects due to redundancy, Aldh1a1/Aldh1a3 double knockouts exhibit severe eye defects during the late stage of morphogenesis (Matt et al. 2005, Molotkov et al. 2006). Aldh1a1/Aldh1a3 double knockouts demonstrated that RA generated in the optic cup is not required locally for development of the retina, nor for eye dorsoventral patterning as previously suggested from RA gain-of-function studies, but instead, RA generated in the dorsal and ventral retina diffuses outside of the optic cup to surrounding perioptic mesenchyme to regulate its migration around the optic cup during anterior eye formation (Matt et al. 2005, Molotkov et al. 2006); in the absence of RA signaling, there is excessive migration of perioptic mesenchyme resulting in ectopic corneal and eyelid tissue that secondarily exerts a mechanical force that distorts optic cup morphology. This late loss of RA signaling in the eye decreases expression of Pitx2 and Foxc1 in perioptic mesenchyme, which are known regulators of eye morphogenesis (Matt et al. 2005). Further studies revealed that after RA activates Pitx2 expression in the perioptic mesenchyme, Pitx2 then activates Dkk2 which functions to inhibit WNT signaling that limits migration of the perioptic mesenchyme (Kumar & Duester 2010).
Perspectives for future studies
Great progress has been made in understanding RA signaling; however, a reproducibility crisis exists (Duester 2017). As mentioned above, particularly with regard to the role of RA signaling during limb development, functions for RA based upon gain-of-function studies are often not supported by in vivo genetic loss-of-function studies likely due to off-target effects induced by RA treatment. Another issue is whether RA signaling may also act nongenomically to control processes other than nuclear gene transcription using a mechanism that involves cytoplasmic RARs that control mRNA translation or other nongenomic processes (Liou et al. 2005, Aoto et al. 2008). However, the nongenomic hypothesis has not been supported by genetic loss-of-function studies that remove endogenous RA, suggesting that pharmacological levels of RA may lead to nongenomic effects while normal levels of endogenous RA do not function nongenomically (Duester 2022). Thus, a focus on in vivo genetic loss-of-function studies is needed to bring consensus on how RA signaling normally functions during development or in adult tissues. Such studies have identified many tissues that require endogenous RA signaling for normal development as described above (Fig. 2). In reality, even though sometimes very difficult, genetic loss-of-function studies are needed to determine the normal function of any gene, protein, or molecule as gain-of-function studies (treatments with molecules/proteins or overexpression of genes) cannot provide this answer.
For future studies, genomic and epigenetic methods will be more effective than analyzing one gene at a time, thus allowing a global identification of RA-regulated genes and associated RAREs required for developmental processes. Identification of RA direct target genes can be achieved by comparing WT and RA knockout embryos (or specific tissues) using RNA-seq and ChIP-seq to identify genes with significant decreases or increases in expression that also have RA-regulated changes in deposition of the gene activation mark H3K27ac or the gene repression mark H3K27me3 located near RAREs (Berenguer et al. 2020). Direct RA target genes identified in this manner can be subjected to knockout studies to determine if they are required for specific developmental processes, and many times such genes will be found to be essential for mediating the pathways used by RA signaling to control development. This basic knowledge will improve strategies to prevent or treat developmental defects.
Declaration of interest
The authors declare that there is no conflict of interest that could be perceived as prejudicing the impartiality of this review.
Funding
This work was funded by the National Institutes of Health (National Eye Institute) grant R01 EY031745 (G D) and the National Institutes of Health (National Institute of Arthritis and Musculoskeletal and Skin Diseases) grant R56 AR067731 (G D).
Data availability statement
No data is provided in this report.
Acknowledgements
A special thanks to Pierre.
References
Abu-Abed S, MacLean G, Fraulob V, Chambon P, Petkovich M & Dollé P 2002 Differential expression of the retinoic acid-metabolizing enzymes CYP26A1 and CYP26B1 during murine organogenesis. Mechanisms of Development 110 173–177. (https://doi.org/10.1016/s0925-4773(0100572-x)
Agarwal P, Wylie JN, Galceran J, Arkhitko O, Li C, Deng C, Grosschedl R & Bruneau BG 2003 Tbx5 is essential for forelimb bud initiation following patterning of the limb field in the mouse embryo. Development 130 623–633. (https://doi.org/10.1242/dev.00191)
Aoto J, Nam CI, Poon MM, Ting P & Chen L 2008 Synaptic signaling by all-trans retinoic acid in homeostatic synaptic plasticity. Neuron 60 308–320. (https://doi.org/10.1016/j.neuron.2008.08.012)
Begemann G, Schilling TF, Rauch GJ, Geisler R & Ingham PW 2001 The zebrafish neckless mutation reveals a requirement for raldh2 in mesodermal signals that pattern the hindbrain. Development 128 3081–3094. (https://doi.org/10.1242/dev.128.16.3081)
Belyaeva OV, Adams MK, Wu L & Kedishvili NY 2017 The antagonistically bifunctional retinoid oxidoreductase complex is required for maintenance of all-trans-retinoic acid homeostasis. Journal of Biological Chemistry 292 5884–5897. (https://doi.org/10.1074/jbc.M117.776914)
Berenguer M, Lancman JJ, Cunningham TJ, Dong PDS & Duester G 2018 Mouse but not zebrafish requires retinoic acid for control of neuromesodermal progenitors and body axis extension. Developmental Biology 441 127–131. (https://doi.org/10.1016/j.ydbio.2018.06.019)
Berenguer M, Meyer KF, Yin J & Duester G 2020 Discovery of genes required for body axis and limb formation by global identification of retinoic acid-regulated epigenetic marks. PLoS Biology 18 e3000719. (https://doi.org/10.1371/journal.pbio.3000719)
Billings SE, Pierzchalski K, Butler Tjaden NE, Pang XY, Trainor PA, Kane MA & Moise AR 2013 The retinaldehyde reductase DHRS3 is essential for preventing the formation of excess retinoic acid during embryonic development. FASEB Journal 27 4877–4889. (https://doi.org/10.1096/fj.13-227967)
Bliss AF 1951 The equilibrium between vitamin A alcohol and aldehyde in the presence of alcohol dehydrogenase. Archives of Biochemistry and Biophysics 31 197–204. (https://doi.org/10.1016/0003-9861(5190206-8)
Cunningham TJ & Duester G 2015 Mechanisms of retinoic acid signalling and its roles in organ and limb development. Nature Reviews: Molecular Cell Biology 16 110–123. (https://doi.org/10.1038/nrm3932)
Cunningham TJ, Zhao X, Sandell LL, Evans SM, Trainor PA & Duester G 2013 Antagonism between retinoic acid and fibroblast growth factor signaling during limb development. Cell Reports 3 1503–1511. (https://doi.org/10.1016/j.celrep.2013.03.036)
Cunningham TJ, Brade T, Sandell LL, Lewandoski M, Trainor PA, Colas A, Mercola M & Duester G 2015 Retinoic acid activity in undifferentiated neural progenitors is sufficient to fulfill Its role in restricting FGF8 expression for somitogenesis. PLoS ONE 10 e0137894. (https://doi.org/10.1371/journal.pone.0137894)
Cunningham TJ, Colas A & Duester G 2016 Early molecular events during retinoic acid induced differentiation of neuromesodermal progenitors. Biology Open 5 1821–1833. (https://doi.org/10.1242/bio.020891)
Diez del Corral R, Olivera-Martinez I, Goriely A, Gale E, Maden M & Storey K 2003 Opposing FGF and retinoid pathways control ventral neural pattern, neuronal differentiation, and segmentation during body axis extension. Neuron 40 65–79. (https://doi.org/10.1016/s0896-6273(0300565-8)
Duester G 2017 Retinoic acid’s reproducible future. Science 358 1395. (https://doi.org/10.1126/science.aar6752)
Duester G 2022 Synaptic plasticity is altered by treatment with pharmacological levels of retinoic acid acting nongenomically however endogenous retinoic acid has not been shown to have nongenomic activity. Journal of Neurological Disorders 10 461. (https://doi.org/10.4172/2329-6895.10.1.475)
Dupé V, Matt N, Garnier JM, Chambon P, Mark M & Ghyselinck NB 2003 A newborn lethal defect due to inactivation of retinaldehyde dehydrogenase type 3 is prevented by maternal retinoic acid treatment. PNAS 100 14036–14041. (https://doi.org/10.1073/pnas.2336223100)
Elder TD & Topper YJ 1962 The oxidation of retinene (vitamin A aldehyde) to vitamin A acid by mammalian steroid-sensitive aldehyde dehydrogenase. Biochimica et Biophysica Acta 64 430–437. (https://doi.org/10.1016/0006-3002(6290300-1)
England S, Batista MF, Mich JK, Chen JK & Lewis KE 2011 Roles of Hedgehog pathway components and retinoic acid signalling in specifying zebrafish ventral spinal cord neurons. Development 138 5121–5134. (https://doi.org/10.1242/dev.066159)
Fan X, Molotkov A, Manabe S-I, Donmoyer CM, Deltour L, Foglio MH, Cuenca AE, Blaner WS, Lipton SA & Duester G 2003 Targeted disruption of Aldh1a1 (Raldh1) provides evidence for a complex mechanism of retinoic acid synthesis in the developing retina. Molecular and Cellular Biology 23 4637–4648. (https://doi.org/10.1128/MCB.23.13.4637-4648.2003)
Feng L, Hernandez RE, Waxman JS, Yelon D & Moens CB 2010 Dhrs3a regulates retinoic acid biosynthesis through a feedback inhibition mechanism. Developmental Biology 338 1–14. (https://doi.org/10.1016/j.ydbio.2009.10.029)
Ghyselinck NB & Duester G 2019 Retinoic acid signaling pathways. Development 146 dev167502. (https://doi.org/10.1242/dev.167502)
Giguère V, Ong ES, Segui P & Evans RM 1987 Identification of a receptor for the morphogen retinoic acid. Nature 330 624–629. (https://doi.org/10.1038/330624a0)
Henrique D, Abranches E, Verrier L & Storey KG 2015 Neuromesodermal progenitors and the making of the spinal cord. Development 142 2864–2875. (https://doi.org/10.1242/dev.119768)
Hernandez RE, Rikhof HA, Bachmann R & Moens CB 2004 vhnf1 integrates global RA patterning and local FGF signals to direct posterior hindbrain development in zebrafish. Development 131 4511–4520. (https://doi.org/10.1242/dev.01297)
Hernandez RE, Putzke AP, Myers JP, Margaretha L & Moens CB 2007 Cyp26 enzymes generate the retinoic acid response pattern necessary for hindbrain development. Development 134 177–187. (https://doi.org/10.1242/dev.02706)
Hoffmann B, Lehmann JM, Zhang XK, Hermann T, Husmann M, Graupner G & Pfahl M 1990 A retinoic acid receptor-specific element controls the retinoic acid receptor-b promoter. Molecular Endocrinology 4 1727–1736. (https://doi.org/10.1210/mend-4-11-1727)
Kastner P, Grondona JM, Mark M, Gansmuller A, LeMeur M, Decimo D, Vonesch JL, Dollé P & Chambon P 1994 Genetic analysis of RXRa developmental function: convergence of RXR and RAR signaling pathways in heart and eye morphogenesis. Cell 78 987–1003. (https://doi.org/10.1016/0092-8674(9490274-7)
Kastner P, Mark M & Chambon P 1995 Nonsteroid nuclear receptors: what are genetic studies telling us about their role in real life? Cell 83 859–869. (https://doi.org/10.1016/0092-8674(9590202-3)
Keegan BR, Feldman JL, Begemann G, Ingham PW & Yelon D 2005 Retinoic acid signaling restricts the cardiac progenitor pool. Science 307 247–249. (https://doi.org/10.1126/science.1101573)
Kochhar DM 1973 Limb development in mouse embryos. I. Analysis of teratogenic effects of retinoic acid. Teratology 7 289–298. (https://doi.org/10.1002/tera.1420070310)
Krumlauf R 1993 Hox genes and pattern formation in the branchial region of the vertebrate head. Trends in Genetics 9 106–112. (https://doi.org/10.1016/0168-9525(9390203-t)
Kumar S & Duester G 2010 Retinoic acid signaling in perioptic mesenchyme represses Wnt signaling via induction of Pitx2 and Dkk2. Developmental Biology 340 67–74. (https://doi.org/10.1016/j.ydbio.2010.01.027)
Kumar S & Duester G 2014 Retinoic acid controls body axis extension by directly repressing FGF8 transcription. Development 141 2972–2977. (https://doi.org/10.1242/dev.112367)
Kumar S, Cunningham TJ & Duester G 2016 Nuclear receptor corepressors Ncor1 and Ncor2 (Smrt) are required for retinoic acid-dependent repression of FGF8 during somitogenesis. Developmental Biology 418 204–215. (https://doi.org/10.1016/j.ydbio.2016.08.005)
Liou JC, Ho SY, Shen MR, Liao YP, Chiu WT & Kang KH 2005 A rapid, nongenomic pathway facilitates the synaptic transmission induced by retinoic acid at the developing synapse. Journal of Cell Science 118 4721–4730. (https://doi.org/10.1242/jcs.02603)
Lohnes D, Mark M, Mendelsohn C, Dollé P, Dierich A, Gorry P, Gansmuller A & Chambon P 1994 Function of the retinoic acid receptors (RARs) during development. (I) Craniofacial and skeletal abnormalities in RAR double mutants. Development 120 2723–2748. (https://doi.org/10.1242/dev.120.10.2723)
Maden M 1982 Vitamin A and pattern formation in the regenerating limb. Nature 295 672–675. (https://doi.org/10.1038/295672a0)
Maden M, Gale E, Kostetskii I & Zile MH 1996 Vitamin A-deficient quail embryos have half a hindbrain and other neural defects. Current Biology 6 417–426. (https://doi.org/10.1016/s0960-9822(0200509-2)
Mariani FV, Ahn CP & Martin GR 2008 Genetic evidence that FGFs have an instructive role in limb proximal-distal patterning. Nature 453 401–405. (https://doi.org/10.1038/nature06876)
Marshall H, Studer M, Pöpperl H, Aparicio S, Kuroiwa A, Brenner S & Krumlauf R 1994 A conserved retinoic acid response element required for early expression of the homeobox gene Hoxb-1. Nature 370 567–571. (https://doi.org/10.1038/370567a0)
Matt N, Dupé V, Garnier JM, Dennefeld C, Chambon P, Mark M & Ghyselinck NB 2005 Retinoic acid-dependent eye morphogenesis is orchestrated by neural crest cells. Development 132 4789–4800. (https://doi.org/10.1242/dev.02031)
Mendelsohn C, Lohnes D, Décimo D, Lufkin T, LeMeur M, Chambon P & Mark M 1994 Function of the retinoic acid receptors (RARs) during development. (I) Craniofacial and skeletal abnormalities in RAR double mutants. Development 120 2749–2771. (https://doi.org/10.1242/dev.120.10.2749)
Mercader N, Leonardo E, Piedra ME, Martínez-A C, Ros MA & Torres M 2000 Opposing RA and FGF signals control proximodistal vertebrate limb development through regulation of Meis genes. Development 127 3961–3970. (https://doi.org/10.1242/dev.127.18.3961)
Mic FA, Haselbeck RJ, Cuenca AE & Duester G 2002 Novel retinoic acid generating activities in the neural tube and heart identified by conditional rescue of Raldh2 null mutant mice. Development 129 2271–2282. (https://doi.org/10.1242/dev.129.9.2271)
Mic FA, Molotkov A, Benbrook DM & Duester G 2003 Retinoid activation of retinoic acid receptor but not retinoid X receptor is sufficient to rescue lethal defect in retinoic acid synthesis. PNAS 100 7135–7140. (https://doi.org/10.1073/pnas.1231422100)
Molotkov A, Molotkova N & Duester G 2006 Retinoic acid guides eye morphogenetic movements via paracrine signaling but is unnecessary for retinal dorsoventral patterning. Development 133 1901–1910. (https://doi.org/10.1242/dev.02328)
Molotkova N, Molotkov A, Sirbu IO & Duester G 2005 Requirement of mesodermal retinoic acid generated by Raldh2 for posterior neural transformation. Mechanisms of Development 122 145–155. (https://doi.org/10.1016/j.mod.2004.10.008)
Niederreither K, Subbarayan V, Dollé P & Chambon P 1999 Embryonic retinoic acid synthesis is essential for early mouse post-implantation development. Nature Genetics 21 444–448. (https://doi.org/10.1038/7788)
Niederreither K, Vermot J, Schuhbaur B, Chambon P & Dollé P 2000 Retinoic acid synthesis and hindbrain patterning in the mouse embryo. Development 127 75–85. (https://doi.org/10.1242/dev.127.1.75)
Niederreither K, Vermot J, Messaddeq N, Schuhbaur B, Chambon P & Dollé P 2001 Embryonic retinoic acid synthesis is essential for heart morphogenesis in the mouse. Development 128 1019–1031. (https://doi.org/10.1242/dev.128.7.1019)
Novitch BG, Wichterle H, Jessell TM & Sockanathan S 2003 A requirement for retinoic acid-mediated transcriptional activation in ventral neural patterning and motor neuron specification. Neuron 40 81–95. (https://doi.org/10.1016/j.neuron.2003.08.006)
Pennimpede T, Cameron DA, MacLean GA, Li H, Abu-Abed S & Petkovich M 2010 The role of CYP26 enzymes in defining appropriate retinoic acid exposure during embryogenesis. Birth Defects Research: Part A, Clinical and Molecular Teratology 88 883–894. (https://doi.org/10.1002/bdra.20709)
Petkovich M, Brand NJ, Krust A & Chambon P 1987 A human retinoic acid receptor which belongs to the family of nuclear receptors. Nature 330 444–450. (https://doi.org/10.1038/330444a0)
Pradhan A, Zeng XI, Sidhwani P, Marques SR, George V, Targoff KL, Chi NC & Yelon D 2017 FGF signaling enforces cardiac chamber identity in the developing ventricle. Development 144 1328–1338. (https://doi.org/10.1242/dev.143719)
Probst S, Kraemer C, Demougin P, Sheth R, Martin GR, Shiratori H, Hamada H, Iber D, Zeller R & Zuniga A 2011 SHH propagates distal limb bud development by enhancing CYP26B1-mediated retinoic acid clearance via AER-FGF signalling. Development 138 1913–1923. (https://doi.org/10.1242/dev.063966)
Rallis C, Bruneau BG, Del Buono J, Seidman CE, Seidman JG, Nissim S, Tabin CJ & Logan MPO 2003 Tbx5 is required for forelimb bud formation and continued outgrowth. Development 130 2741–2751. (https://doi.org/10.1242/dev.00473)
Rhinn M & Dolle P 2012 Retinoic acid signalling during development. Development 139 843–858. (https://doi.org/10.1242/dev.065938)
Ryckebusch L, Wang Z, Bertrand N, Lin SC, Chi X, Schwartz R, Zaffran S & Niederreither K 2008 Retinoic acid deficiency alters second heart field formation. PNAS 105 2913–2918. (https://doi.org/10.1073/pnas.0712344105)
Sandell LL, Sanderson BW, Moiseyev G, Johnson T, Mushegian A, Young K, Rey JP, Ma JX, Staehling-Hampton K & Trainor PA 2007 RDH10 is essential for synthesis of embryonic retinoic acid and is required for limb, craniofacial, and organ development. Genes and Development 21 1113–1124. (https://doi.org/10.1101/gad.1533407)
Sirbu IO, Gresh L, Barra J & Duester G 2005 Shifting boundaries of retinoic acid activity control hindbrain segmental gene expression. Development 132 2611–2622. (https://doi.org/10.1242/dev.01845)
Sirbu IO, Zhao X & Duester G 2008 Retinoic acid controls heart anteroposterior patterning by down-regulating Isl1 through the FGF8 pathway. Developmental Dynamics 237 1627–1635. (https://doi.org/10.1002/dvdy.21570)
Sorrell MR & Waxman JS 2011 Restraint of FGF8 signaling by retinoic acid signaling is required for proper heart and forelimb formation. Developmental Biology 358 44–55. (https://doi.org/10.1016/j.ydbio.2011.07.022)
Stratford T, Horton C & Maden M 1996 Retinoic acid is required for the initiation of outgrowth in the chick limb bud. Current Biology 6 1124–1133. (https://doi.org/10.1016/s0960-9822(0270679-9)
Studer M, Pöpperl H, Marshall H, Kuroiwa A & Krumlauf R 1994 Role of a conserved retinoic acid response element in rhombomere restriction of Hoxb-1. Science 265 1728–1732. (https://doi.org/10.1126/science.7916164)
Tickle C, Alberts BM, Wolpert L & Lee J 1982 Local application of retinoic acid to the limb bud mimics the action of the polarizing region. Nature 296 564–566. (https://doi.org/10.1038/296564a0)
White JC, Shankar VN, Highland M, Epstein ML, DeLuca HF & Clagett-Dame M 1998 Defects in embryonic hindbrain development and fetal resorption resulting from vitamin A deficiency in the rat are prevented by feeding pharmacological levels of all-trans-retinoic acid. PNAS 95 13459–13464. (https://doi.org/10.1073/pnas.95.23.13459)
Wilson JG, Roth CB & Warkany J 1953 An analysis of the syndrome of malformations induced by maternal vitamin A deficiency. Effects of restoration of vitamin A at various times during gestation. American Journal of Anatomy 92 189–217. (https://doi.org/10.1002/aja.1000920202)
Yashiro K, Zhao X, Uehara M, Yamashita K, Nishijima M, Nishino J, Saijoh Y, Sakai Y & Hamada H 2004 Regulation of retinoic acid distribution is required for proximodistal patterning and outgrowth of the developing limb. Developmental Cell 6 411–422. (https://doi.org/10.1016/s1534-5807(0400062-0)
Zhao X, Sirbu IO, Mic FA, Molotkova N, Molotkov A, Kumar S & Duester G 2009 Retinoic acid promotes limb induction through effects on body axis extension but is unnecessary for limb patterning. Current Biology 19 1050–1057. (https://doi.org/10.1016/j.cub.2009.04.059)
